the
united nations
educational, scientific
and cultural
organization
ab<]US
i
international centre for theoretical physics
international atomic
energy agency
H4.SMR/1304-7
"COLLEGE ON SOIL PHYSICS"
12 March - 6 April 2001
Lecture Notes: Soil Physics
R. Harttnann
University of Ghent
Laboratory of Soil Physics
Dept. Soil Management & Soil Care
Faculty of Agriculture & Applied Biological Sciences
Ghent, Belgium
These notes are for internal distribution onlv
strada costiera, I I - 34014 trieste italy - tel.+39 04022401 I I fax +39 040224163 - scLinfo@ictp.trieste.it - www.ictp.trieste.it
UNIVERSITHT
GENT
LECTURE NOTES : SOIL PHYSICS
by
R. HARTMANN1
1
Laboratory of Soil Physics
Department of Soil Management and Soil Care
Faculty of Agricultural and Applied Biological Sciences
University Gent - Belgium
TEXTBOOKS
Rose, C.W., 1956.
Agricultural physics.
Slayter,R.O., 1967.
Plant-water relationships.
New York: Academic.
Childs,E.C, 1969.
An introduction to the physical basis of soil water phenomena.
New York: Intersciences.
Hillel,D., 1971.
Soil and water: Physical properties and processes.
New York: Academic.
Nielsen, D.R., Jackson, R.D., Corry, J.W. and Evans, D.D., 1972.
Soil water.
Madison, Wisconsin, Am. Soc. of Agronomy.
Baver, L.D., Gardner, W.H. and Gardner, W.R., 1972.
Soil Physics.
New York: Wiley.
Kirkham, D. and Powers, W.C., 1972.
Advanced soil physics.
New York: Interscience.
Taylor, S.A. and Ashcroft, G.L., 1972.
Physical edaphology.
San Francisco: W.H. Freeman and Company.
Drainage Principles and Applications, 1973.
I. Introductory subjects.
II. Theories of field drainage and watershed runoff.
III. Surveys and investigations.
IV. Design and management of drainage systems.
International Institute for Land Reclamation and Improvement, Wageningen,
The Netherlands.
Hillel,D., 1974.
L'eau et sol.
Principes et processus physiques.
Louvain: Vander.
Yong, R.N. and Warketin, B.P, 1975.
Soil properties and behaviour.
Amsterdam: Elsevier.
Henin?S.?1977.
Cours de physique du sol, I et II.
Paris: Orstom.
Marshall, TJ. and Holmes, J.W., 1979.
Soil physics.
Cambridge: Cambridge University Press.
Hanks, R.J. and Ashcroft, G.L., 1980.
Applied soil physics.
Berlin: Springer-Verlag.
Hillel,D., 1980.
Fundamentals of soil physics.
New York: Academic.
Hillel,D., 1980.
Applications of soil physics.
New York: Academic.
Koorevaar, P., Menelik, G. and Dirksen, C, 1983.
Elements of Soil Physics.
Elsevier, Amsterdam.
Campbell, G., 1985.
Soil physics with basic.
Elsevier, Amsterdam.
Iwata, S., Tabuchi, T. and Warketin, 1988.
Soil-water interactions. Mechanisms and applications.
New-York and Basel: Marcel Dekker, Inc.
Ghildyal, B.P. and Tripathi, R.D., 1987.
Soil Physics.
John Wiley and Sons, New-York.
Jury, W.A., Gardner, W.R. and Gardner, W.H, 1991.
Soil Physics.
Fifth Edition: John Wiley & Sons, Inc.
Smith, K.A. and Mullins, C.E., 1991.
Soil Analysis (Physical Methods).
Marcel Dekker Inc., New York, Bazel, Hong Kong.
Topp, G.C., Reynolds, W.D. and Gran, R.E., 1992.
Advances in measurement of soil physical properties:
Bringing theory into practice.
SSSA Special Publication n° 30.
Hanks, R.J., 1992.
Applied Soil Physics.
Soil Water and Temperature Application, 2th edition, Springer-Verlag.
Kutilek, M. and Nielsen, D.R., 1994.
Soil Hydrology
Geo-ecology textbook, Catena Verlag.
TABLE OF CONTENTS
1. SOIL WATER CONTENT
2
THERMOGRAVIMETRIC METHOD
6
2. SOIL WATER POTENTIAL
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
INTRODUCTION
ENERGY STATE OF SOIL WATER
QUANTITATIVE EXPRESSION OF SOIL WATER POTENTIAL
GRAVITATIONAL POTENTIAL
OSMOTIC POTENTIAL
MATRIC POTENTIAL
EXTERNAL GAS PRESSURE POTENTIAL
HYDRAULIC HEAD
3. TENSIOMETER
3.1
3.2
3.3
3.4
3.5
HYDROSTATIC PRESSURE POTENTIAL - PIEZOMETER
MATRIC POTENTIAL - TENSIOMETER
PRINCIPLE OF THE TENSIOMETER
HOW TO CALCULATE THE SOIL WATER PRESSURE HEAD H AND THE HYDRAULIC HEAD H
SOME CHARACTERISTICS OF THE TENSIOMETER
3.5.1 Cup conductance "C"
3.5.2 Sensitivity of the manometer "S"
3.6 PRACTICES AND LIMITATIONS OF TENSIOMETERS
3.7 APPLICATIONS OF MEASUREMENTS
7
7
7
10
10
12
13
16
17
20
20
21
22
24
30
30
30
32
34
3.7.1 Determination of the direction of water flow at different levels in the soil profile
(figure 18)
34
3.7.2 Flux control at a certain depth
35
3.7.3 Determination of the soil water characteristic curve (or retentivity curve)
35
3.7.4 Scheduling irrigation
38
4. SOIL WATER CHARACTERISTIC CURVE
4.1 MEASUREMENT
41
41
4.1.1 Hanging Water Column (Range -100 cm < h < 0)
4.1.2 Pressure Plate (Range - 15000 cm < h < - 300 cm)
4.1.3 Equilibration over Salt Solutions (h < - 15000 cm)
41
42
43
4.2 HYSTERESIS IN WATER CONTENT-ENERGY RELATIONSHIPS
44
5. HYDRAULIC CONDUCTIVITY OF SATURATED AND UNSATURATED SOIL 48
5.1. REPRESENTATIVE ELEMENTARY VOLUME
5.2. SATURATED HYDRAULIC CONDUCTIVITY ' K S '
5.2.1 Laboratory methods
5.2.2 Field methods
5.2.3 Models
5.3. UNSATURATED HYDRAULIC CONDUCTIVITY K(9)/K(H)
5.3.1 Steady state flow
5.3.2 Non steady state/ transient flow
48
50
50
50
50
50
50
50
5.3.3 Models
50
5.4. MODELS FOR DETERMINATION OF THE SOIL WATER CHARACTERISTIC CURVE AND THE
UNSATURATED HYDRAULIC CONDUCTIVITY K ( 9 )
51
5.4.1 Introduction
5.4.2 Fitting pF curve to a discrete set of measuring points by modelling
5.4.2.1 The model of van Genuchten
51
51
51
5.4.2.LI Principle
51
5.4.2.1.2 Graphical interpretation and physical meaning of the parameters
52
5.4.2.2 Other models
54
5.4.2.2.1 The model of Brooks and Corey
54
5.4.2.2.2 The model of Campbell
54
5.4.3 Methods to obtain the pF curve without laboratory determinations
54
5.4.3.1 Soil texture reference curves
54
5.4.3.2 Estimating soil water retention curves from soil physical characteristics
55
5.4.3.2.1 Introduction
55
5.4.3.2.2 Estimation of specific points on the soil water retention curve
56
5.4.3.2.3 Estimation of soil water retention model parameters
56
5.4.4 Determination of the unsaturated hydraulic conductivity K(9) by modelling
58
5.4.4.1 Introduction
58
5.4.4.2 Estimating the unsaturated hydraulic conductivity K(9) from water retention data ...58
5.4.4.2.1 Introduction
58
5.4.4.2.2 Estimating K(9) based on the model ofChilds and Collis-George
60
5.4.4.2.3 Estimating K(0) based on the model of Burdine
61
5.4.4.2.4 Estimating K(Q) based on the model of Mualem
61
5.4.4.3 Estimating the unsaturated hydraulic conductivity K(9) from soil physical
characteristics
62
5.4.4.3.1 Soil texture reference curves
62
5.4.4.3.2 Pedotransfer functions to estimate K(9)
63
5.4.5 References
64
Introduction
In soil physics more than in most subjects there has been a strong symbiotic relationship
between the development of theory and practical applications on the one hand and the
development of experimental methods on the other. For example, acceptance and use of the
concept of matric potential, originally outlined by Buckingham in 1907, were greatly
accelerated following Richards' description in 1942 of simple methods for constructing
reliable tensiometers. The development of the neutron probe, which allows rapid
nondestructive field measurement of soil water content, has led to an enormous upsurge of
interest in soil water measurements and their practical applications and to the accelerated
development of the theory of soil water movement. In comparison with soil chemistry and
mineralogy, the development of techniques for soil physical measurements has been held
back by the necessity for many measurements to be carried out in the field.
Other techniques that require reliable 'undisturbed' cores or samples on which to make
laboratory measurements have also required specification of a number of soil physical
measurement techniques, which is currently being undertaken by the Soil Quality working
group of the International Standards Organisation, is an indication of both the increasing
importance of soil physical measurements and the coming of age of a number of physical
measurement methods. These methods can be regarded as being relatively stable, in the sense
that their limitations, reliability, and areas of application are now well established.
Despite these current attempts at standardisation, the choice of appropriate soil physical
measurement techniques is often still uncharted ground full of pitfalls. Abandoned and
unused soil physical equipment in research laboratories provides a sorry testimony to the
difficulties faced by researchers who have been unable to obtain adequate advice on choice of
appropriate methods. There are certainly many books, reviews, and technical notes that
provide listings and descriptions of soil physical methods. However, most users require a
brief guide to any necessary theory and to the techniques most suitable to their own particular
application before they can make a choice and are ready to consult one of the many available
complications of methods.
Recent books written to provide all the information required to allow the user to choose the
technique most suited to his or her desired application are given in the reference list. These
books contain mainly a critical review of relevant theory and measurement methods.
Particular emphasis has been given to the merits, limitations, and range of application of each
method. As well as the consideration of accuracy, this includes measurement time, case of
use, and cost.
1. SOIL WATER CONTENT
Soil water is important in itself as a feature of the physical environment but especially prominent
in its relationships with climatology and with the surface and the subsurface hydrologic regimes
as a component in the terrestrial water balance. To understand the behaviour of soil water one
must measure it. This have long proved a difficult task both from the instrumental aspect as
because of the complexity of the soil body surpasses that of the vegetated layer and the
atmosphere above.
There are numerous procedures and types for the determination of the soil water content. Since
each method has its advantages and limitations it is well to consider both the purpose for which
determinations are to be made and the features of each possible method including the cost of
buying equipment and operating and maintenance costs.
Eight of these are listed in table 1. Only the most frequently applied methods will be discussed.
Table 1. Methods of measuring soil water content
1. Thermogravimetric method
2. Neutron scattering technique
3. Gamma-ray attenuation technique
4. Gamma-ray backscattering technique
5. Electrical resistance
6. Thermal conductivity
7. Time Domain Reflectometry
8. Capacitance method
The soil water content (wetness) can be expressed in terms of either mass or volume ratios or
fractions.
- Mass wetness or dry mass fraction of water (w).
w
=
Ms
This is the mass of water relative to the mass of dry (105°C) soil particles where w is the mass
wetness, Mw water mass and Ms dry (105°C) soil mass.
-
Volume wetness or volume fraction of water (6).
V
l
+ V
It is the dimensionless ratio of the water volume (Vw) relative to total bulk soil volume Vt. The
latter is the sum of the volume of solids (Vs), water (Vw) and air (Va) (figure 1).
The two expressions can be related to each other as follows:
9
w
with
Vw Ms
V, Mw
(3)
: Ms/Vt = pb (mass of dry soil per unit bulk volume and usually lies in the range of
1.3 and 1.7 gem"3)
Mw/Vw = pw (mass of water per unit volume of water and is approximately equal to
1 g cm"3)
Equation (3) becomes
w ph
S\
'
D
So in order to obtain the soil water content on a volume basis the mass wetness is multiplied
with the dry bulk density. Both w and 0 are usually multiplied by 100 and reported as
percentages by mass or volume.
The usefulness of the soil water content by volume lies in the fact that it can be converted easily
into head units of water therefore being compatible with quantities of rainfall or irrigation water
applied. All calculations involving the water balance of the soil, including calculation of
irrigation deficits, water application efficiency and recharge of the soil moisture reservoir by
rainfall, involve the use of the volumetric soil water percentage (vol.% of water = mm of water
per 10 cm soil depth).
If the soil water content is measured at different depths of the profile than the depth-interval dzj
for which Q\ is valid, is taken as the vertical distance between the measuring point and the points
located half-way respectively the above and underlying measuring point (figure 2).
The soil water storage "S" is than equal to the equivalent waterhead (mm or cm) present in the
profile until a depth z and is obtained through the summation of the moisture contents over each
depth interval:
s = LQ
d
?-i
(5)
i=7
with the same unit as that of z which is compatible with the amount of precipitation or
evaporation.
Two moisture content profiles, measured at respectively time ti and t2, allows us to calculate the
change in water storage AS(zj, zi) during the period At.
volume
relation
mass
iciauu 11
Air
/
A
/
>
M a = ;o
/ \
>
/V
' w
>
V
m
Ms
y
Vt
Water
siiiliiiliHii
USB
Jl|S
ilfc
ii
>
Figure 1. Schematic diagram of the soil as a three-phase system.
soil moisture content 9
dz
t:
S (z,^
N
Figure 2. Soil moisture content profiles.
Thermogravimetric method
The thermogravimetric method of measuring soil water content consists of removing a sample
by augering into the soil and then determining its moist and dry weight. The moist weight is
determined by weighing the sample as it is at the time of sampling and the dry weight is
obtained after drying the sample to a constant weight in an oven. The standard method of drying
is to place the sample in an oven at 105°C for 24 hours.
The mass wetness is the ratio of the weight loss in drying to the dry weight of the sample:
w =
wet weight - dry weight
dry weight
weight loss by drying
weight of dried sample
The method depending at it does on sampling, transporting and repeated weighings entails
practically inevitable errors. It is also laborious and time consuming, since a period of at least 24
hours is usually considered necessary for complete oven drying. The standard method of oven
drying is itself arbitrary. Some clays may still contain appreciable amounts of adsorbed water
even at 105°C. On the other hand, some organic matter may oxidize and decompose at this
temperature so that the weight loss may not be due entirely to the vaporization of water.
The errors of the gravimetric method can be reduced by increasing the sizes and the number of
samples. However the sampling method is destructive and may disturb an observation or
experimental plot sufficiently to distort the results. For these reasons many prefer indirect
methods, which permit making frequent or continuous measurements at the same point and once
the equipment is installed and calibrated with much less time, labor and soil disturbance.
Although the gravimetric determination of soil moisture content is rather laborious it is, because
of its simplicity and reliability, the most extensively applied technique and is used as calibration
standard for other methods.
2. SOIL WATER POTENTIAL
2.1 Introduction
Soil water content is not sufficient to specify the entire status of water in soil. For example, if
soils with a same water content but with different particle size distribution are placed in
contact with each other, water will flow from a coarse textured soil to a fine textured soil.
One needs to define a property that will help to explain this observation.
Perhaps the following analogy will help. Heat content (analogous to soil water content) is a
property of a material that is useful for many purposes. It will not, however, tell us directly
whether heat will flow. Therefore a heat intensity term, temperature, has been defined which
permits to determine the direction of heat flow. The soil water term that is analogous to
temperature (i.e. the intensity with which the water is in the soil) is called the soil water
potential. Water potential is a much more complicated property than temperature.
2.2 Energy state of soil water
Soil water, like other bodies in nature, can contain energy in different quantities and forms.
Classical physics recognizes two principal forms of energy, kinetic and potential. Since the
movement of water in the soil is quite slow, its kinetic energy, which is proportional to the
velocity squared, is generally considered to be negligible. On the other hand, the potential
energy, which is due to position or internal condition, is of primary importance in
determining the state and movement of water in the soil.
The potential energy of soil water varies over a very wide range. Differences in potential
energy of water between one point and another give rise to the tendency of water to flow
within the soil. The spontaneous and universal tendency of all matter in nature is to move
from where the potential energy is higher to where it is lower and to equilibrate with its
surroundings. In the soil, water moves constantly in the direction of decreasing potential
energy until equilibrium, definable as a condition of uniform potential energy throughout, is
reached.
The rate of decrease of potential energy with distance is in fact the moving force causing
flow. A knowledge of the relative potential energy state of soil water at each point within the
soil can allow us to evaluate the forces acting on soil water in all directions, and to determine
how far the water in a soil system is from equilibrium. This is analogous to the well-known
fact that an object will tend to fall spontaneously from a higher to a lower elevation, but that
lifting it requires work. Since potential energy is a measure to the amount of work a body can
perform by virtue of the energy stored in it, knowing the potential energy state of water in the
soil and in the plant growing in that soil can help us to estimate how much work the plant
must expend to extract a unit amount of water.
Clearly, it is not the absolute amount of potential energy "contained" in the water which is
important in itself, but rather the relative level of that energy in different regions within the
soil. The concept of soil water potential is a criterion, for this energy. It expresses the specific
potential energy (= per unit mass) of soil water relative to that of water in a standard reference
state. The standard state generally used is that of a hypothetical reservoir of pure free water
(i.e. water not influenced by the solid phase), at atmospheric pressure, at the same
temperature as that of soil water (or at any other specified temperature) and at a given and
constant elevation.
It is the convention to assign to free and pure liquid water a potential value of zero.
Since the elevation of this hypothetical reservoir can be set at will, it follows that the potential
which is determined by comparison with this standard is not absolute, but by employing even
so arbitrary a criterion we can determine the relative magnitude of the specific potential
energy of water at different locations or times within the soil.
The concept of soil water potential is of great fundamental importance. This concept replaces
the arbitrary categorizations which prevailed in the early stages of the development of soil
physics and which purported to recognize and classify different forms of soil water : e.g.
gravitational water, capillary water, hygroscope water.
New definition by the soil physics terminology committee of the International Soil Science
Society provided more clarity in what used to be a rather complicated theoretical set of
criteria. The total potential of soil water was defined as follows : " the amount of work that
must be done per unit quantity (mass, volume or weight) of pure free water in order to
transport reversibly and isothermally an infinitesimal quantity of water from a pool of pure
water at a specified elevation at atmospheric pressure (standard reference state) to the soil
water at the point under consideration in the soil-plant-atmosphere-system" (figure 3).
If work is required the potential is positive, but if water in the reference state can accomplish
work in moving into the soil the potential is negative.
Soil water is subjected to a number of force field which cause its potential to differ from that
of pure free water. Such forces result from the attraction of the solid matrix for water, as well
as from the presence of dissolved salts and the action of the local pressure in the soil gas
phase and the action of the gravitational field. Accordingly the total potential (v|/t) of soil
water relative to a chosen standard rate can be thought of as the sum of the separate
contributions of the various components as follows :
where
:
i|/t
v|/g
\\i0
v|/m
v|/e<p
...
=
=
=
=
=
=
total soil water potential
gravitational potential
osmotic potential
matric potential
external gas pressure potential
additional terms are theoretically possible
#C
(atmosphere)
(plant)
soil surface =
reference level
pure water
standard state Wt = 0
Figure 3. Potential of soil water, water in plant cell and water in the atmosphere.
The main advantage of the total potential concept is that it provides a unified measure by
which the state of water can be evaluated at any time and every where within the soil-plantatmosphere system.
2.3 Quantitative expression of soil water potential
The dimensions of the soil water potential are those of energy per unit quantity of water and
the units depend on the way the quantity is specified. Common alternatives used are :
a. Energy per unit mass of water (7/kg)
This method of expression is not widely used.
b. Energy per unit volume of water (pressure) (J/m3 or N/m2)
This is the most common method of expressing potential and can be written with units of
either Pascal or bar or atmosphere.
c. Energy per unit weight of water (head) (J/N = Nm/N = m)
This method of expressing potential is also common and has units of length.
For conversion from one unit to another knows that:
1 bar corresponds to 100 J/kg
lbar-105Pa
1 bar corresponds to 10 m water head
2.4 Gravitational potential
Every body on the earth's surface is attracted towards the centre of the earth by a
gravitational force equal to the weight of the body, that weight being the product of the
body's mass by the gravitational acceleration. To rise a body against this attraction, work
must be expended and this work is stored by the rised body in the form of gravitational
potential energy. The amount of this energy depends on the body's position in the
gravitational force field.
The gravitational potential of soil water at each point is determined by the elevation of the
point relative to some arbitrary reference level. If the point in question is above the reference,
v|/g is positive. If the point in question is below the reference, \|/g is negative. Thus the
gravitational potential is independent of soil properties. It depends only on the vertical
distance between the reference and the point in question.
10
Conversion table for units of soil water potential (*)
Volumetric potential
units
Specific potential
units
(*)
Weight potential
units
Joule kg"]
bar
millibar
Pa
atmosphere
0.0001
0.000001
0.001
0.1
0.000000987
1
0.01
10
103
0.00987
10.17 xlO" 2
100
1
1000
105
0.987
10.17
102
0.000987
1.017 x l O 2
m
0.001017 x l O 2
0.1
0.001
1
101.3
1.013
1013
1.013 xlO 5
1
10.30
0.09833
0.0009833
0.9833
98.33
0.0009703
io- 2
The density of water was taken as 1.000 g cm"3. This holds only at 4°C but is approximately correct at other temperatures.
At a height z below a reference level (e.g. the soil surface) the gravitational potential of a
mass M of water, occupying a volume V is :
- Mgz=-pwVgz
where
:
pw =
g =
density of water
acceleration of gravity
Gravitational potential can be expressed :
per unit mass
:
ty%~-?>z
per unit volume :
y/^ = y/g pw = - pw g z
per unit weight
:
(J/kg)
(Pa)
¥ew=—=~
g
2.5 Osmotic potential
The osmotic potential is attributable to the presence of solutes in the soil water. The solutes
lower the potential energy of the soil water. Indeed, the fact that water molecules move
through a semi-permeable membrane from the pure free water into a solution (osmosis)
indicates that the presence of solutes reduces the potential energy of the water on the solution
side (figure 4). At equilibrium sufficient water has passed through the membrane to bring
about significant difference in the heights of liquid. The difference (z) in the levels represents
the osmotic potential.
Since the osmotic potential of pure free water is zero the osmotic potential of a solution at the
same temperature of free water is negative (water flow occurs from point of high potential to
one with lower potential).
Differences in osmotic potential only play a role in causing movement of water when there is
an effective barrier for salt movement between the two locations at which the difference in v|/0
was observed. Otherwise, the concentration of salts will become the same throughout the
profiles by the process of diffusion and the difference in v|/0 will no longer exist. Therefore
osmotic potential does not act as a driving force in water flux. This potential is of importance
in water movement into and through plant roots, in which there are layers of cells which
exhibit different permeabilities to solvent and solute.
12
osmotic
pressure
salt solution
pure water
*>
semi-permeable
membrane
Figure 4. Schematic presentation of osmosis.
2.6 Matric potential
Matric potential results from forces associated with the colloidal matric and includes forces
associated with adsorption and capillarity. These forces attract and bind water in the soil and
lower its potential energy below that of bulk water. The capillarity results from the surface
tension of water and its contact angle with the solid particles. In an unsaturated (three-phase)
soil system, curved menisci form which obey the equation of capillarity :
where
:
?{
=
pressure of soil water, can be smaller than atmospheric
Pa
=
atmospheric pressure, conventionally taken as zero
AP
=
pressure deficit
y
=
surface tension of water
R1? R2
=
principal radii of curvature of a point on the meniscus,
taken as negative when the meniscus is concave
As we assume the soil pores to have a cylindrical shape (figure 5) the meniscus has the same
curvature in all directions and equation above becomes :
R
13
since :
R=-
r
cosa
(a = 0 : angle of contact between water and the soil particle surface)
Pf = AP =
where
:
r
h
pw
g
=
=
=
=
with AP equals - h p w g
radius of the capillary tube
height of capillary rise
density of water (10 3 kg/m 3 )
acceleration of gravity (9.81 m/s 2 « 10 m/s 2 « 10 N/kg)
If the soil were like a simple bundle of capillary tubes, the equations of capillarity might be
themselves suffice to describe the relation of the negative pressure potential or matric
potential to the radii of the soil pores in which the menisci are contained. However, in
addition to the capillarity phenomenon, the soil also exhibits adsorption, which forms
hydration envelopes, over the particle surfaces. These two mechanisms of soil water
interaction are illustrated in figure 6.
The presence of water in films as well as under concave menisci is most important in clayey
soil and at high suctions or low potential, and it is influenced by the electric double layer and
the exchangeable cations present. In sandy soils adsorption is relatively unimportant and the
capillary effect predominates.
In general, however, the matric potential results from the combined effect of the two
mechanisms, which cannot easily be separated since the capillary "wedges" are at a state of
internal equilibrium with the adsorption "films" and the ones cannot be changed without
affecting the others. Hence matric potential denotes the total effect resulting from the affinity
of the water of the whole matric of the soil, including its pores and particle surfaces together.
14
\
\
subatmospherifc^
pressure
h2
water
negative pressure
potential
positive pressure
3otential
superatmospheric
pressure
capillary tube
free water
surface
Figure 5. Capillary rise of water into a capillary tube.
"adsorbed" water
particles
Figure 6. Water in an unsaturated soil is subject to capillarity and adsorption, which combine
to produce a matric potential.
15
The matric potential can be expressed :
- per unit mass
:
y/ - -gh =
(J/kg)
- per unit volume
:
y/m pw = - p w gh =
- per unit weight
:
1
-2y
y/ — = -h =
g
Pwgr
-2/
(Pa)
(m)
The matric potential is a dynamic property of the soil.
In saturated soil (below the ground water level) the liquid phase is at hydrostatic pressure
greater than atmospheric and thus its pressure potential is considered positive (figure 5). Thus
water under a free water surface is at a positive pressure potential (hydrostatic pressure
potential \|/h), while water at such a surface is at zero pressure potential (assuming
atmospheric pressure in the soil) and water which has risen in a capillary tube above that
surface is characterized by a negative pressure or matric potential.
Since soil water may exhibit either of the two potentials, but not both simultaneously, the
matric and the hydrostatic pressure potential are referred to as the pressure potential (v|/p).
Nevertheless it is an advantage in unifying the matric potential and hydrostatic pressure
potential in that this unified concept allows one to consider the entire profile in the field in
terms of a single unsaturated zone, below and above the water table.
2.7 External gas pressure potential
A factor which may affect the pressure of soil water is a possible change in the pressure of the
ambient air. In general this effect is negligible in the field as the atmospheric pressure
remains nearly constant small barometric pressure fluctuations notwithstanding. However, in
the laboratory the application of excess air pressure to change the soil water pressure is a
common practice resulting into the so called external gas pressure of pneumatic potential.
16
FINAL REMARKS
1. Matric potential and the former term matric suction are numerically equal - when
expressed in the same units - but except for the sign.
2. The effect of an external gas pressure different from the atmospheric (reference) pressure
is generally also included in the pressure potential so that:
Accordingly the total potential being :
characterizes fully the state of water in soil under the prevailing conditions ; the gradients
of these three parameters are the basis for transport theory.
2.8 Hydraulic head
The total potential is obtained by combining the relevant component potentials
Equilibrium, which is defined as the situation where mass transfer of water in the liquid phase
is absent, is obtained when the value of the total potential at different points in the system in
constant. Usually, sufficient condition is that the sum of the component potentials, \\JQ being
ignored, is constant. The equilibrium condition states then that:
y/g + y/p= cons tan t = y/H
(6)
called hydraulic potential.
As already stated, the external gas pressure or pneumatic potential in the field may be
assumed to be zero. Also the soil water within a profile may exhibit either matric or
hydrostatic pressure potential (figure 5) but not simultaneously. Therefore it is an advantage
in unifying both in a single continuous potential extending from the saturated region into the
unsaturated region below and above the water table.
17
As it is often usual to designate the potential in terms of head, equation (6) becomes
H=h+z
where
: H = the hydraulic head (m)
h = the soil water pressure head (m)
> 0 under the water table (saturated zone)
< 0 above the water table (unsaturated zone)
z = the gravitational head (m)
The definition is very important because the hydraulic gradient between two points under
consideration in a soil is the driving force for water movement.
In figure 7 the condition is applied to a vertical soil column in equilibrium with a water table.
No water movement occurs in the column. The water table is taken as the reference level for
the gravitational potential.
Under the water table matric potential equals zero, but a pressure potential called hydrostatic
pressure potential occurs which can also be presented by a value of h but with always a
positive sign.
18
VH(H)
0.2
0.4
soil moisture
content (v/v)
ground water level
reference level
height (cm)
\\i (z,cm)
30
30
20
20
10
10
0 reference level 0
-10
-10
-20
-20
(h,cm)
(h,cm)
-30
-20
-10
0
0
0
0
0
0
0
10
20
Figure 7. Equilibrium condition in a soil column.
19
(H,cm)
0
0
0
0
0
0
3. TENSIOMETER
3.1 Hydrostatic pressure potential - Piezometer
As discussed earlier the hydrostatic (positive) pressure potential \j/h under field conditions
applies to saturated soils and is measured with a piezometer (figure 8).
A piezometer is a tube of a few cm inner diameter, open at both ends, which is installed in a
soil profile. If the lower end is below the groundwater table, a piezometer is partially filled
with water. By determining the height of the water level in a piezometer it is possible to
calculate the (positive) hydrostatic pressure potential of the soil water at the lower end of the
tube. The diameter of piezometer is chosen large enough that capillary rise and resistance to
water flow are negligible. As a result, any variation in hydraulic potential that may arise
inside the piezometer, is instantaneously equalized. Thus, even if the hydrostatic pressure
potential at the lower end is changing rapidly, the water inside a piezometer goes through a
series of static equilibria and at any moment it can be assumed that the hydraulic head is
uniform and equal to the hydraulic head of the soil water at the open lower end. There
exchange of water takes place such that the pressure is always locally uniform. The static
hydraulic head in piezometer can be determined by measuring the depth of the water level,
since at the flat air-water interface the pressure potential is zero.
soil surface
0.5 m
L ground water level
0.3 m
Figure 8. A piezometer in a soil profile.
20
Figure 8 shows a piezometer in a soil profile in which the water is at static equilibrium. The
reference point z = 0 is taken at the soil surface. In the piezometer at the water level H = h + z
= 0 - 0.5 m = - 0.5 m. Thus at point A, H must also be - 0.5 m (H = h + z = 0.30 m - 0.80 m =
- 0.50 m).
The hydrostatic pressure potential expressed per unit weight of water at any point in the soil
under the water table is the distance between the point and the water level in the piezometer
tube.
The water level in a piezometer tube is at the level of the groundwater table in a situation of
static equilibrium, independent of the depth of the lower end.
3.2 Matric potential - Tensiometer
Piezometers cannot be used to measure negative pressure potentials because in unsaturated
conditions, water flows out of the tube into the soil leaving the tube dry. The negative
pressure or matric potential can be measured with the so-called tensiometer.
The tensiometer consists of a liquid filled porous cup, mostly of ceramic material and
connected to a pressure measuring device such as a mercury manometer or vacuum gauge via
a liquid-filled tube (figure 9).
1. Flexible water reservoir
2. Gauge to measure tension
3. Soil level
4. Filled with water
5. Porous end (cup) through which water can move
6. Soil
Figure 9. Tensiometer with a vacuum gauge (jet fill tensiometer).
21
If the ceramic cup is embedded in soil, the soil solution can flow into or out of the
tensiometer through the very small pores in the ceramic cup. Analogously to the situation
discussed for piezometer, this flow continues until the (negative) pressure potential of the
liquid in the cup has become equal to the (negative) pressure potential of the soil water
around the cup. Thus the (negative) pressure potential called matric potential v|/m of soil water
can be measured with a tensiometer, and is therefore also often called tensiometer pressure
potential.
3.3 Principle of the tensiometer
When the cup is placed in a water reservoir (figure 10), the water inside the cup comes into
hydraulic contact with the water in the reservoir through the water-filled small pores in the
ceramic walls. The water level in the tube will indicate the level of the water in the reservoir.
,B
water
/
porous cup
Figure 10.Porous cup connected with a piezometer tube for measuring pressure potentials
under the water table.
The pressure is given by the height h of the water level above the middle of the porous cup
and the pressure PA equals :
where
=
=
density of water
acceleration due to gravity
22
If we place now the porous cup, connected with an U-shape water filled tube in a soil than the
bulk water inside the cup will come in hydraulic contact with the liquid phase in the soil.
When initially placed in the soil, the water in the tensiometer is at atmospheric pressure. Soil
water in unsaturated soil has a negative pressure and therefore exercises a suction which
drawn out a certain amount of water from the rigid and air-tight tensiometer, causing a drop
in the water level at the open end of the U-tube (figure 11).
/
/
soil surface
-A
porous cup
'7*
B'
-water
Figure 11 .Tensiometer for measuring pressure potentials in soils.
The drier the soil, the higher the suction and the lower the water level at equilibrium in the
U-tube. The height h of the liquid column that has moved into ("sucked into") the soil in
figure 11 is therefore an index of the magnitude of the potential, or :
As h is measured downwards the minus sign is introduced so that PA gives a negative
pressure.
This type of tensiometer is very simple and useful to illustrate the basic principles involved.
Practical applications often do not allow the use of the water manometer because the U-tube
extends below the level of the tensiometer cup and measurements thus requires inconvenient,
deep pits. Therefore open manometers, filled with immiscible liquids of different densities
such as mercury are used so that these problems do not arise (figure 12).
23
Using mercury implies that a relatively short height indicated a relatively large pressure
difference in the manometer (1 cm of mercury corresponds to 13.55 cm of water). Besides the
simple water or mercury manometer a vacuum gauge or an electrical transducer is also used.
1. stop
r\
^
2. water
3. mercury
4. soil
5. porous cup
\
Figure 12. Tensiometer with mercury manometer.
3.4 How to calculate the soil water pressure head h and the hydraulic head H
Let x be the height of the mercury in the manometer (cm) and z the vertical axis (figure 13).
At the interface water-mercury in the manometer, the pressure is the same in water and in
mercury (being PB). The repartition of the pressure is hydrostatic in the water column
between point B and the tensiometer cup (point A), but also between point B and the free
surface of the mercury in the reservoir (point C).
Using the hydrostatic law for liquids in equilibrium one obtains per unit weight of liquid the
following hydraulic head equation :
P
z+
^constant
Pg
where
:
z
P/p g
=
=
gravitational head
pressure head
24
water
B
x
reference level
A
Figure 13. Tensiometer installation with the mercury level in the reservoir at the soil surface
being the reference level.
From figure 13 one obtains :
in water
P
P
=z
g
_l_
P§
Because soil surface is taken as reference level for the gravitational potential, and point A is
located below that level, the gravitational head is negative (- zA).
(7)
25
:
p
p.
zB + — — = z c + c
zc =
Pc =
reference level = 0
atmospheric pressure = 0
in mercury
because :
the equation becomes :
or :
B~
(8)inC;0 gives
since
:
=0
(8)
Pfig g^B
:
Pw
=
PH B
=
PA
~ Pfig g ZB
Pwg
Pwg
(9)
1000 kg m"3
13600 kg m"3
(9) becomes:
or
PB
zB +
hA =-13.6
z +z"AA +zB
'•"•* ~B B
J
:
hA=-12.6zB+zA
hA=-12.6x+zA
(10)
Normally the free surface of the mercury in the reservoir (point C) is located y cm above the
soil surface (reference level) (figure 14).
Equation (10) becomes
:
hA=-12.6x+y+zA
(11)
The hydraulic head H, being the sum of the pressure head h and the gravitational head z,
(H = h + z), becomes :
H=-12.6x+y + zA+(-zA)
H=-12.6x + y
26
(12)
water
B
Hg
/
/
/
/
/
/
A
reference level
A
Figure 14. Tensiometer installation with the mercury reservoir y cm above soil surface.
If the soil water pressure head around the porous cup changes, the height of the mercury level
will change consequently. For a situation where e.g. two tensiometers are connected to the
same mercury reservoir, the value y is the same and from equation (12) it follows that:
Hl-H2=-l2.6(xt-x2)
This means that tensiometers connected to the same mercury reservoir yield hydraulic head
differences irrespective of there depths. So H} - H2 = 0 where xx = x2. Such installation
provides at once the direction of the water flow: from the tensiometers with low mercury
height to those with high mercury height.
27
The mercury manometer can be replaced by a Bourdon or dial vacuum gauge (figure 15) or
by a pressure transducer (figure 16). For such manometers equation (11) and equation (12)
became :
removable cap
0—vacuum gauge
soil surface
tube
porous wall cup
free water
"Tn "porous cup
soil aggregates (including
particles, air and water)
pores
porous cup
ENLARGED VIEW
Figure 15. Principle of tensiometer with vacuum gauge.
28
soil surface
1. nitril-caoutchouc septum stopper
2. plexi-glass tube (lucid)
3. air pocket (VJ (end of syringe needle here)
4. deaerated water
5. ceramic cup
MEASURING DEVICE
1. air vent
2. containment with p = p (atm)
3. semiconductor element
4. steel membrane (D = 13 mm)
5. guiding tube
6. guiding tube for needle
7. syringe needle (D = 0.4 mm) dead volume (V())
8. guiding spring
9. membrane chamber (45 mm)
10. shielded four lead wire
Figure 16.Diagram of tensiometer with septum stopper and pressure transducer with attached
syringe needle.
29
3.5 Some characteristics of the tensiometer
3.5.1 Cup conductance "C"
Being the volume of water dV that has to be displaced per unit time between the tensiometer
and the soil through the porous cup for the measuring device to register the change in soil
water pressure head dh :
(m2/s)
The surface area, conductivity and wall thickness of the porous cup can be varied to increase
its conductance and consequently reduce the measuring time. However there are practical
limits to this e.g. the air entry value of the porous cup.
3.5.2 Sensitivity of the manometer "S"
The sensitivity of the manometer determines the performance of the pressure measuring
device :
where dV is the volume of water displaced needed by the measuring device to register a
pressure head change dh.
• water/mercury manometer
The water manometer has the lowest sensitivity as it requires a large volume of water to be
displaced per unit soil water pressure head change. If the diameter of the manometer tube is
large enough to eliminate capillary effects, soil water pressure head can be measured with a
precision of ± 1 mm. Since large change in soil water pressure head requires an exchange of a
large volume of water with the soil it is convenient to replace water by mercury.
30
The mercury manometer reduces the height of the liquid column and thus the displaced
volume of water by a factor 12.6 (equation 11) and so increases the sensitivity by the same
factor. The overall precision of mercury manometers is often not better than 5 cm WH. Also
the size of the manometer tube influences the sensitivity. Indeed if the cross-sectional area of
a manometer tube is A cm2 the water displacement head AV is A Ah so that:
S=
Ah 1
=—
A Ah A
2
y(in J)
• Bourdon/dial vacuum gauges
The sensitivity are generally comparable to that of the mercury manometer.
• Pressure transducers
They have very high sensitivity since they require very small displacements of their sensing
element to register a full scale pressure range. However the sensitivity of the tensiometer
system is different from that of the pressure sensor. The volume of water that needs to be
exchanged with the soil is generally larger because the tubing and air present in the system
expand and compress with changes in pressure.
• Response time "T,"
It is the time needed to make an accurate measurement with a particular tensiometer system
and is inversely proportional to the conductance of the porous cup of the tensiometer and the
sensitivity of the pressure measuring device and is characterized as follows :
r
( )
This is for a system were the tensiometer is the limiting factor (in water or wet soil). In a soillimited system (dry soil) the response is much slower under decreasing soil water pressure
heads due to small water flow rates to the tensiometer cup as a result of low soil hydraulic
conductivities. The flexibility of the tensiometer tubing and the compressibility of air bubbles
inside the tubing decrease the effective sensitivity and thus increase the effective response
time of a tensiometer system.
31
3.6 Practices and limitations of tensiometers
The purpose of the measurements with tensiometers is to characterize the existing pressure
potential of the soil water.
Water within the tensiometer should be continuous throughout the system to allow a correct
transfer of pressure from the soil to the mercury. Occurrence of gas bubbles disrupts this
continuity and makes the system inoperative. The fine porous cup has the function of not
allowing penetration of air from the unsaturated soil into the water-filled tensiometer tube,
even though water can and should move through it. The fine pores inside the wall of the
ceramic cup have a high air-entry value which is the pressure needed to remove the water
from the pores in the cup replacing it by air. Even with a high air entry value breakdown of
the system occurs due to entrapped air within the tensiometer tube or to air coming out of
solution at reduced pressure.
Due to the fact that the manometer measures a partial vacuum relative to the external
atmospheric pressure, measurements by tensiometry are generally limited-to about - 850 cm
of waterhead. Use of tensiometers in the field is therefore only possible when pressures do
not fall below this value. Moreover the pressure head in the tensiometer must stay above the
air-entry value of the porous material which is gives as follows :
h=
where
:
-A a cos 6
—
PwSdmax
h
= pressure head at which air enters the cup through the largest pore
a
= surface tension of an air-water interface (0.0073 N m'1 at 25°C)
9
= contact angle
dmax = largest pore diameter
So to cover the practical tensiometer range to h = - 850 cm for a completely wetting material
(0 = 0) dmax must be smaller than 3.6 \xm. If the air-entry value is exceeded air can enter the
tensiometer and the soil will drain the tensiometer.
However, the limited range of pressure measurable by the tensiometers is not as serious as it
may seen at first sight. In many agricultural soils the tensiometer range accounts for more
than 50% of the amount of soil water taken up by the plants. To what extend the available
water range expressed e.g. as a percentage of the water between pF 2 and pF 4.2 is covered
by the tensiometer depends on the shape of the moisture characteristic curve (pF-curve) as
shown for three soil types in figure 17.
32
Thus where soil management (particularly in irrigation) is aimed at maintaining high pressure
potential conditions which are mostly favourable for plant growth, tensiometers are definitely
useful.
Air diffusion through the porous cup into the system requires frequent purging with deaired
water. Tensiometers are also sensitive to temperature gradients between their various parts.
Hence the above-ground parts should preferably be shielded from direct exposure to the sun.
Therefore it is also suggested to make readings always at the same time of the day (e.g. at
08.00 a.m.).
available water range
tensiometer range
20
30
40
50
vol °/o
water
Figure 17. Part of the available moisture range covered by tensiometers, depending on soil
type.
1. Sand 50% of available moisture
2. Loam 75% of available moisture
3. Clay 25% of available moisture
When installing a tensiometer it is important for proper functioning that good contact be
made between the porous cup and the surrounding soil. Generally the porous cup is pushed
into a hole with a slightly smaller diameter to ensure good contact. If the soil is initially rather
dry and hard, prewetting of the hole may be necessary. In a stony soil a small excavation
should be made and filled with very fine sand into which the tensiometer can be placed.
33
With mercury manometers, even when small diameter nylon tubing (+/- 2 mm) is used, often
a considerable volume of water must be adsorbed by the soil (during water uptake or drying
process) or by the porous cup (replenishing by rainfall or irrigation) before the potential that
really exists can be read off correctly. A very convenient modern device, the electronic
transducer can be used which reacts to very small changes in pressure and converts these
changes in a small electrical current which can be registered and amplified by a voltmeter.
This system is very accurate but also very sensitive to the occurrence of small air bubbles in
the tensiometer system. Moreover is it rather expensive.
Since the porous cup walls of the tensiometer are permeable to both water and solutes, the
water inside the tensiometer tends to assume the same solute composition and concentration
as soil water, and the instrument does not indicate the osmotic potential of soil water.
3.7 Applications of measurements
3.7.1
Determination of the direction of water flow at different levels in the soil profile
(figure 18)
The concept of the water potential is well suited for the analysis of water flow in soils, since
all flow is a consequence of potential gradients. Darcy's law, though originally conceived for
saturated flow only, was extended to unsaturated flow, with provision that conductivity is a
function of soil water content 0.
For a vertical one dimensional water flow Darcy's equation can be written as follows :
dH
q = -K(0)—
where
:
q
=
K(0) =
H
=
=
flux
hydraulic conductivity
hydraulic head
h + zwith h = soil water pressure head
z = gravitational head
34
(13)
The minus sign in the equation indicates that the flow is in the direction of decreasing potential.
This means also that if we have two tensiometers located at depths zl and z2 (zl < z2) :
- q will be negative (upward flow - evaporation) if H2 > H^ the rise of mercury in
manometer nr 2 is lower than in manometer nr 1
- q will be positive (downward flow - percolation) if H2 < H^ the opposite situation is
observed
- q will be zero (plane of zero flux) at a certain depth z when the curve H(z) will show a
maximum or the rise of the mercury a minimum because dH/dz = 0. A graphical example
is presented in figure 18
3.7.2 Flux control at a certain depth
From agricultural point of view it could be of interest to know it there is a recharge of the
water table or capillary rise. Therefore only 2 tensiometers are needed with a depth distance
of say 25 cm in the control zone. A simple reading of the rise of mercury in the manometer
will indicate the flow direction.
Knowing the moisture content 0 at the depth between zx and z2? the K(0) relation of that soil
and the hydraulic head gradient dH/dz, one can calculate the instantaneous water flow q (see
equation 13).
3.7.3 Determination of the soil water characteristic curve (or retentivity curve)
The h-9 relation (retentivity curve) of a soil layer in situ can be established :
- knowing the soil water pressure head (h) using tensiometers (see equation 11)
- knowing the soil water content (9) using the neutron moisture meter
35
reference level
0
0
0
0
0
) = zA + y - 1 2 . 6 x
)
H (cm)^
• 200
H
= -12.6x
-100
1
dz
-- 100
0
downward flux
(percolation)
z(cm)
Figure 18. Hydraulic head profiles. The manometers from left to right increase with depth.
36
- 200
H
-100
dz
- 100
upward flux
(evaporation)
(5)"
z(cm)
H (cm)«
fdz < 0 , q > 0
z(cm)
Figure 18bis. Hydraulic head profiles. The manometers from left to right increase with depth.
37
3.7.4 Scheduling irrigation
The root zone for most agricultural plants is limited to the unsaturated part of the profile
because the plant roots do not proliferate in a saturated soil where aeration is limiting.
Consequently in a non-saline soil the plant behaviour is largely determined by the matric
potential of the soil water. Moreover the plant does not depend as much on the quantity of
water present as it does on the water potential.
Water should be applied to the soil when the matric potential is still high enough that the soil
can and does supply water fast enough to meet the atmospheric demands without placing the
plant under a stress that will reduce yield or quality of the harvested crop.
Although the tensiometers function over only a limited part of the available water range (0 to
- 800 cm water) it is usually in this range that plants should be irrigated.
From practical point of view tensiometers are installed at minimum 2 locations. One unit
should be placed in the zone of maximum root activity and another near the bottom of the
active rootzone.
The time to irrigate is determined by following the matric potential readings in the zone of
the greatest root activity. The exact value of the matric potential at which water should be
applied is not the same for every crop. A good approximation of that matric potential is
available for many common crops. For most crops it is time to irrigate when the top
tensiometer reads - 300 to - 500 cm water and the bottom tensiometer begins to indicate
drying (Table 1).
38
Table 1. Matric potential at which water should be applied for maximum yields of various
crops grown in deep, well-drained soil that is fertilized and otherwise managed for
maximum production. Where two values are given, the higher value is used when
evaporative demand is high and the lower value when it is low; intermediate values
are used when the atmospheric demand for evapotranspiration is intermediate. (The
values are subject to revision as additional experimental data become available).
(TAYLOR and ASHCROFT, 1972)
Crop
Vegetative crops
Alfalfa
Beans (snap and lima)
Cabbage
Canning peas
Celery
Grass
Lettuce
Tobacco
Sugar cane
Tensiometer
Blocks
Sweet corn
Turfgrass
Root crops
Onions
Early growth
Bulbing time
Sugar beets
Potatoes
Carrots
Broccoli
Early
After budding
Cauliflower
Matric potential
(joules/kg)
Equivalent matric suction
(centibars)
-150
-75 to -200
-60 to -70
-30 to -50
-20 to -30
-30 to-100
-40 to -60
-30 to -80
150
75 to 200
60 to 70
30 to 50
20 to 30
30 to -00
40 to 60
30 to 80
-15 to-50
-100 to-200
-50 to-100
-24 to -36
15 to 50
100 to 200
50 to 100
24 to 36
-45 to -55
-55 to -65
-40 to -60
-30 to -50
-55 to -65
45 to 55
55 to 65
40 to 60
30 to 50
55 to 65
-45 to -55
-60 to -70
-60 to -70
45 to 55
60 to 70
60 to 70
39
Crop
Fruit crops
Lemons
Oranges
Deciduous fruit
Avocadoes
Grapes
Early season
During maturity
Strawberries
Cantaloupe
Tomatoes
Bananas
Grain crops
Corn
Vegetative period
During ripening
Small grains
Vegetative period
During ripening
Seed crops
Alfalfa
Prior to bloom
During bloom
During ripening
Carrots
During seed year at
60 cm depth
Onions
During seed year
at 7 cm depth
at 15 cm depth
Lettuce
During productive
phase
Coffee
Matric potential
(joules/kg)
Equivalent matric suction
(centibars)
-40
-20 to-100
-50 to -80
-50
40
20 to 100
50 to 80
50
-40 to -50
<-100
-20 to -30
-35 to -40
-80 to-150
-30 to-150
40 to 50
>100
20 to 30
35 to 40
80 to 150
30 to 150
-50
-800 to-1200
50
800 to 1200
-40 to -50
-800 to-1200
40 to 50
800 to 1200
-200
-400 to -800
-800 to-1500
200
400 to 800
800 to 1500
-400 to -600
400 to 600
-400 to -600
400 to 600
-150
150
-300
300
Requires short periods of low potential to break bud dormancy,
followed by high water potential
40
4. SOIL WATER CHARACTERISTIC CURVE
Text from "Soil Physics" by Jury W.R, Gardner W.R. and Gardner W.H. (John Wiley &
Sons, Inc., New York)
4.1 Measurement
In rigid porous media, the matric potential as defined in the preceding represents the effect of
adsorptive soil solid forces and interfacial curvature on water potential energy. The functional
relationship between the matric potential and the gravimetric or volumetric water content is
called the water characteristic function or matric potential-water content function v|/m (0). This
function may be evaluated by measuring matric potential and water content simultaneously
with the methods already discussed during a succession of water content changes. In the
laboratory, v|/m (9) may be measured on replicated prepared samples over a large range of
water contents. Virtually the entire range from water-saturated soil to very dry soil may be
covered by using a hanging water column, a pressure membrane, and equilibration over salt
solutions. These devices will be illustrated using the equilibrium principle.
4.1.1 Hanging Water Column (Range -100 cm < h < 0)
A hanging water column consists of a water-saturated, highly permeable porous ceramic plate
connected on its underside to a water column terminating in a reservoir open to the
atmosphere. Water-saturated samples of soil held in rings are placed in contact with the flat
plate when the water reservoir height is even with the top of the plate. Then the reservoir is
lowered to a new height a distance z below the top of the plate (figure 19).
soil sample
porous ceramic plate
free water surface
water column
ll
I
overflow
Figure 19.Desaturation of soil water samples to a desired energy state with a hanging water
column.
41
By the equilibrium principle, water will flow from the soil samples through the ceramic to the
reservoir until the total water potential of the system is constant. At this time the potential of
the free reservoir may be set equal to zero, and at the soil sample height z we may write
(z = 0; P = Patm; neglect solutes) v|/m + \j/z = 0 = i|/m + pwgz, or i|/m = - pwgz. When equilibrium
has been restored, some of the samples may be removed and their gravimetric or volumetric
water content measured. The tube may then be lowered further and a new set of samples
measured. If there is good contact between the soil and the ceramic, equilibrium will be
reached rapidly (i.e. several hours) since the samples are quite moist. The range of the device
is limited chiefly by the space available for lowering the water column.
4.1.2 Pressure Plate (Range - 15000 cm < h < - 300 cm)
The pressure plate consists of an air-tight chamber enclosing a water-saturated, porous
ceramic plate connected on its underside to a tube that extends through the chamber to the
open air. Saturated soil samples are enclosed in rings and placed in contact with the ceramic
on the top side. The chamber is then pressurized, which squeezes water out of the soil pores,
through the ceramic, and out the tube (figure 20). At equilibrium, flow through the tube will
cease. We may set the total potential equal to zero at the point where the water exits the tube.
Inside the plate we may write (P = Patm; z = 0; neglect solutes) i[/ep + vj/m = 0 = \\im + AP, or
\|/m = - AP. When equilibrium is reached, the chamber may be depressurized and the water
content of the samples measured. An assumption is made in this method that the matric
potential of the sample does not change as the air pressure is lowered to atmospheric.
pressure chamber
soil sample
porous ceramic plate
water-filled tube
Figure 20.Desaturation of soil water samples to a desired energy state with a pressure plate.
42
This method may be used up to air gauge pressures of about 15 bars if special fine-pore
ceramic plates are used. Since these devices have a very high flow resistance, it may require a
substantial amount of time to remove the last small amount of water from the soil. Thus, the
time of equilibrium is difficult to estimate.
4.1.3 Equilibration over Salt Solutions (h < - 15000 cm)
By adding precalibrated amounts of certain salts, the energy level of a reservoir of pure water
may be lowered to any specified level. If this reservoir is brought into contact with a moist
soil sample, water will flow from the sample to the reservoir. If the sample and the reservoir
are placed adjacent to each other in a closed chamber at constant temperature, water will be
exchanged through the vapor phase by evaporation from the soil sample and condensation in
the reservoir until equilibrium is reached. Since the reservoir is a pool of salt solution, at
equilibrium the total potential will be v|/t = vj/so of the solution. In the soil \j/t = i|/m + i|/0 since
the air-water interface acts as a solute membrane. Thus v|/m = v|/so -1|/ 0 , the difference between
the solute potentials of the reservoir and the soil. In practice, the soil will usually not be saline
enough for its solute potential to be significant compared to v|/0 in the range where these
measurements are made. The equilibration time for this method can be shortened by creating
a partial vacuum in the chamber. Care should be taken that the sample and the reservoir are at
the same temperature, because even small temperature differences will cause the soil and salt
solution to equilibrate at very different potentials. Figure 21 shows typical matric potentialvolumetric water content curves for a sandy soil and a finer textured soil high in clay
measured from soil initially at water saturation. The water characteristic function for a soil
desorbed from saturation may be roughly divided into three regions, as shown in the figure.
The air entry region corresponds to the region at saturation where the matric potential
changes but the water content does not. The minimum suction that must be applied to a
saturated soil to remove water from the largest pores is called the air entry suction, which
varies from about 5 to 10 cm for sands to much higher values in unaggregated, fine-textured
soils. After air begins to enter the system, incremental increases in suction on the soil sample
will drain progressively smaller pores, and the water content will drop. This intermediate part
of the curve is called the capillary region. When essentially all of the water held in pores has
been drained, only the tightly bound water adsorbed to particle surfaces remains. Large
changes in matric potentials in this region, called the adsorption region, are associated with
small changes in water content. The differences in the shapes of the water characteristic
function for the prototype sandy and clay soils in figure 21 may be explained by considering
the properties of the bulk solid phases. The clay soil generally has a lower bulk density and
hence a higher water content at saturation. The clay soil has very few large pores and a broad
distribution of particle sizes. Hence, it decreases gradually in water content with decreases in
matric potential. The sandy soil, on the other hand, has much of its water held in large pores
that drain at modest suctions. Hence it will have a very rapid decrease in water content in the
43
capillary region. Finally, the clay soil has a very large surface area compared to the sand and
will have a large amount of water adsorbed to the surfaces.
CLAY SOIL
0.1
0.2
0.3
0.4
0.5
0.6
volumetric water content Gv
Figure 21.Matric potential-water content function (water characteristic function).
4.2 Hysteresis in Water Content-Energy Relationships
Water content and the potential energy of soil water are not uniquely related because the
potential energy state is determined by conditions at the air-water interfaces and the nature of
surface films rather than by the quantity of water present in pores. Soil pores are highly
variable in size and shape and interconnect with each other in a variety of ways. Common to
porous media are so-called bottleneck pores, which have large cavities but narrow points of
connection to adjacent pores. Water is held most tenaciously in small pores, which fill first
when water is admitted to a system. But they do not always empty again during drying in the
same order as they were filled. The factors involved in hysteresis may be discussed most
clearly by assuming that the soil is initially completely devoid of water and subsequently has
no air phase (just liquid water and its vapor). If water is added at this point to the system,
small pores fill first, followed by successively larger and larger pores until all pores are filled
and the matric potential is zero. At intermediate values of saturation, with enough water in the
system so that vapor-water interfaces can exist between particles and in small pores, the
curvature of such interfaces is given by AP = - 2y/R where the pressure difference AP refers to
44
the difference in pressure between the vapor and the liquid water. Water content and water
potential in such a system will follow the wetting curve in figure 22. Some small pores could
be isolated during wetting, so that they might remain dry while larger pores are filled.
However, this would not be the case at equilibrium in the absence of air, in as much as vapor
transfer would assure the wetting of all pores small enough to retain water at a particular
matric potential.
Water content
Scanning
Figure 22.Wetting and drying curves and scanning curves.
If the system is dried either by evaporating water or by bringing the soil into contact with a
dry, porous material that pulls water away from the system, pores will begin to empty,
generally from large to small. However, liquid water may now be trapped in large pores in
such a way that they will not empty in the order that they filled. Water will be held in large
pores until conditions are reached where at least one interconnecting smaller pore can empty;
at this time the larger pore quickly empties. The sudden release of a relatively large amount
of water from a large pore floods surrounding pores and increases the matric potential in them
temporarily. If matric potential were monitored in a small porous system having discrete
differences in pore size, the matric potential-water content relationship for drying might be
saw-toothed, as indicated by the drying curve in figure 22. In real soil systems, the pore size
distribution contains many pores in all size ranges, and the water content and potential
distributions tend to average out so that a smooth curve is obtained. However, the water
content for a given matric potential is higher than for the wetting system, as is shown by the
drying curve in figure 22. This principle is illustrated by the pore-water system in figure 23a.
Here it may be observed that the curvature of the vapor-water interface in the small pores of
two identical systems can be in equilibrium with each other even though their water contents
45
are grossly different. The matric potential is determined by the curvature of the liquid
interface which at equilibrium would be precisely the same in the small pores connecting with
the large pore in each case. This ideal representation commonly is called the "ink bottle
principle", which refers to the fact that an ink bottle has a small opening into a large cavity.
Large pores that are interconnected by smaller pores are not required for hysteresis to occur.
It is possible for a single pore to contain the same amount of water at two different water
potentials, as is shown in figure 23b. Water that condenses initially into such a capillary from
a humid environment is shown by the diagonal cross-hatched area. However, as condensation
proceeds, water at the center finally coalesces and a concave meniscus is formed as a
consequence of surface tension forces (vertical cross-hatching). Whereas positive pressure
existed in the system before coalescence, the system suddenly goes under negative pressure
as a consequence of its new configuration. In the example in figure 23b, water in a cylindrical
pore about 1 cm in length and 0.1 cm in radius would have a slight positive potential of about
0.15 mbar immediately before coalescence and a potential of - 1.5 mbars immediately
afterward.
Surface wetting can also induce hysteresis. Unless particle surfaces are meticulously clean,
they will form a nonzero contact angle with water when wetted (figure 23c). This results in
thicker films than would be present in the drying phase where water films are drawn tightly
over the surface by adsorptive forces.
Thus far the discussion of hysteresis has not involved the presence of air in the system, which
can introduce additional differences between water content at a given matric potential during
wetting and drying. As small pores and interstices between particles fill with water, air may
become entrapped in large pores. Continued water entry into such pores will cause a buildup
of air pressure. Since air is slightly soluble in water, pressures in such pores may gradually be
relieved, which will sometimes allow more complete pore filling. However, the order of
filling and access to pores will be influenced by entrapped air, so that water content is still
permanently affected despite the fact that some air may go into solution and disappear.
In the absence of air, the water potential-water content relationship for complete wetting and
complete drying will follow approximately the dashed line-solid line loop shown in figure 22.
However, this loop is not exactly reproducible because of the inherent difficulty associated
with repetition of the exact order of pore filling over each cycle. When air is present in the
system, the curves are offset somewhat toward the dry side (solid curve, figure 22). If a soil is
completely wetted so that no air is present and then is dried, it will follow the dashed-solid
curve down; upon rewetting in the presence of air, it will follow the solid wetting curve and
will not return to the starting point because of the presence of entrapped air. If the process is
reversed at any time during wetting or drying, curves like those in the interior (dotted curves)
of the hysteretic envelope are produced. These interior curves have been called scanning
curves; the curves that form the hysteretic envelope have been called characteristic curves or
46
the soil moisture characteristic. The wetting curve of the hysteretic envelope is commonly
known as a sorption curve and the drying curve a desorption curve.
Hysteretic phenomena also exist in soil materials as a consequence of shrinking and swelling,
which can affect microscopic pore size geometry as well as overall bulk density. Both factors
would lead to a volumetric water content for a given energy state that differs from that which
would exist if the soil matrix remained fixed. Shrinking and swelling often take place slowly
and usually irreversibly, particularly when organic matter is involved; this complicates the
evaluation of their contribution to hysteresis. Experimental observations do not always reveal
that measurements involve true hysteresis and permanent or semipermanent changes in the
porous system.
(a)
(b)
(c)
Figure 23.Diagrammatic representation of three forms of water content-matric potential
hysteresis.
47
5. HYDRAULIC
CONDUCTIVITY
UNSATURATED SOIL
OF
SATURATED
AND
5.1. Representative elementary volume
The movement of water through soils takes place in the tortuous channels between the soil
particles with velocities varying from point to point. Darcy's law does not consider this
microscopic flow pattern between the particles but instead assumes the water movement to
take place in a continuum with a uniform flow averaged over space. It therefore describes the
flow of water macroscopically in volumes of soil much larger than the size of the pores. In
fact it can be used to describe only the macroscopic flow of water through soil regions of
volume greater than some representative elementary volume that encompasses many soil
particles.
The concept of representative elementary volume of a porous material is most easily
illustrated by considering the measurement of the water content of a sample of unstructured
'uniform' saturated soil, starting with a very small volume and then increasing the sample
size. For very small volumes comparable to the size of the soil particles, the sample volume
would include only solid matter if located wholly within a soil particle, giving zero soil-water
content. If located wholly in a pore, it would contain only water, giving a soil-water content
of one. All values between zero and one are possible when the sample is located partly within
a soil particle and partly within the pore. As the volume size is increased, the lower limit of
measured water content increases while the upper limit decreases, as shown in figure 24.
When the sample is large enough, repeated measurements on random samples of the soil give
the same value of soil-water content. The smallest sample volume that produces a consistent
value is the representative elementary volume. Measurements of hydraulic conductivity and
other soil properties must be made on volumes larger than this volume. While additive soil
properties, such as the water content, can be obtained by averaging a large number of
measurements made on smaller volumes within the representative elementary volume, the
hydraulic conductivity cannot be obtained in this way because of the interdependent complex
pattern of flows that this property embraces.
Sample volume ••
Figure 24.Measurement of soil-water content of a saturated 'uniform' soil (Smith and
Mullins, 1991).
48
Figure 24 illustrates the variability of a soil physical property that exists in all porous
materials at a small enough scale because of their particulate nature. Variability can also be
present in soils at larger scales. For example, in aggregated and structured soils where a
distribution of macropores between the aggregates or peds is superimposed on the
interparticle micropore space, the soil-water content would vary with sample size as shown in
figure 25: only when the sample size encompasses a representative sample of macropore
space do we have a representative volume. This volume will be characteristic of the soil's
structure that determines the hydraulic conductivity.
1.0VoJumetric
soil-water
content
j microscopic
r.e.v.
macroscopic
• r,e.v.
Sample volume
-
Figure 25. Measurement of soil-water content of a saturated soil with superimposed
macro structure (r.e.v. = representative elementary volume) (Smith and Mullins,
1991).
It is only in materials that show behaviour similar to that depicted in figure 24 that continuum
physics, such as that implied by Darcy's law, can be applied macroscopically without
difficulty to soil-water flow phenomena. In materials such as that illustrated in figure 6,
boundary conditions at the surfaces of the aggregates and ever, for saturated conditions, as
long as sufficiently large volumes are considered, continuum physics can still be applied to
water flows at this larger scale using an appropriate value of hydraulic conductivity.
49
5.2. Saturated hydraulic conductivity 6Ks'
5.2.1
Laboratory methods
- constant water head
- variable water head
5.2.2
Field methods
- saturated zone
• auger hole method
• piezometer method
- unsaturated zone
• double ring method
• Guelph/disk permeameter
5.2.3
Models
- capillary tube model
- pedotransferfunctions
5.3. Unsaturated hydraulic conductivity K(9)/K(h)
5.3.1
Steady state flow
- laboratory method
• standard method
• crust method
• sprinkler method
- field method
• crust method
• sprinkler method
5.3.2
Non steady state/ transient flow
- laboratory method : evaporation method of wind
- field method
• internal drainage method
• water balance method
5.3.3
Models
- capillary tube model
- model of Mualem
50
5.4. Models for determination of the soil water characteristic curve and the
unsaturated hydraulic conductivity K(0)
5.4.1
Introduction
The soil moisture characteristic curve (or pF curve) can be determined based on a set of
discrete measuring points determined in the laboratory or in the field. To represent the curve
graphically, one can draw a best fitting curve by hand on mm paper or an analytical equation
can be used to fit a curve through the set of measuring points. Because determination of the
soil moisture characteristic curve in the lab or in the field is time consuming, models have
been introduced to calculate the soil moisture characteristic curve based on readily available
physical soil characteristics.
As regards the unsaturated hydraulic conductivity K(6), can be determined in the laboratory
(e.g. Wind method) or in the field (e.g. internal-drainage method). Since this is also quite
laborious, models have been developed that calculate K(0) based on the saturated hydraulic
conductivity Ks and the soil moisture characteristic curve, or on readily available physical soil
characteristics.
5.4.2
Fitting pF curve to a discrete set of measuring points by modelling
5.4.2.1 The model of van Genuchten
5.4.2.1.1
Principle
There exists a non-linear relationship between the soil moisture content and de matric
potential. To fit a curve through a discrete set of measuring points closed-form analytical
equations can be used. One of the most widely used equations is that of van Genuchten
(1978, 1980) :
/
-er)
V
J
(a-\h\)\
where 0 — volumetric moisture content (m3-m~3),
6r = residual volumetric moisture content (m3-m"3),
6S = volumetric moisture content at saturation (m -m"),
h = matric potential (dimensionless, but expressed in cm
a,nenrn
= parameters (dimensionless).
51
(14)
If, as proposed by van Genuchten,
m =1
for 7i > 1
(15)
n
equation (15) contains 4 independent parameters (6n 9S, a and n), which have to be estimated
for the observed soil-moisture retention data. This can be done by applying a non-linear leastsquares analysis according to the Marquardt (1963) algorithm. This is an iterative method
implying an initial estimate of the parameters.
The Marquardt algorithm can be performed by using mathematical software programmes
such as Mathcad®. On the other hand most statistical software packages, as e.g. SPSS®
contain the Marquardt algorithm. Graphical software such as SigmaPlot® also have the
possibility to perform a non-linear regression. Finally, the RETC code, written in the
programme language Fortran, can be used to determine the parameters of the van Genuchten
model (van Genuchten et al., 1991; Yates et al., 1992).
5.4.2.1.2
Graphical interpretation and physical meaning of the parameters
As stated above the model of van Genuchten contains, after imposing the restriction
m = l-l/n, four independent parameters, namely 0r, 0s, a and n. The value for the volumetric
moisture content at saturation 0s is, in theory, equal to the porosity e of the soil. It is the
latter value that will be taken as initial estimate for 0s.
From a practical point of view the residual volumetric moisture content 0r can be defined
as the pressure head at some large negative value, e.g. at the permanent wilting point
(h =-15,495 cm H2O or pF = 4.2). In some cases, however, significant decreases on h are
likely to result in further desorption of water, especially on fine-textured soils. It seems that
such further changes in #are fairly unimportant for most practical field problems.
To obtain estimates of the remaining parameters a and n a characteristic point P on the curve
has to be defined (see figure 26). This point is located halfway between 6r and 9S. From
equation (14) it follows that the parameter a is inversely proportional with the pressure head
at P, hP :
1
\—
n
1
(16)
a = hp
For n large enough, and hence m approaches 1, a is approximately equal to the inverse value
of hp. Most soils, however, have a m significant smaller than 1 (n < 10). The value for a then
increases exponentially with decreasing m value.
52
2 -
0.1
000
0.2
0.3
0.4
0.5
moisture content (m3.m-3)
data
fitted van Genuchten
Figure.26 pF curve of a heavy sandy clay soil
Finally, the parameter n can be obtained be differentiation of equation (14) (dd-dh'1) and by
substituting equation (16). This gives a relation between h • dd-dh'1 and n. Further calculating
results in following equation for n :
(0 < SP < 1)
n =
,-l
0,5755
1 Sp
(17)
0,025]
(Sp > 1)
where Sp = slope at point P.
The slope Sp is defined as :
SP =
1
es-er
d(log/i)
(18)
From equation (17) and (18) it is clear that as the ratio d#d(log h)~ increases, and hence the
flatter the curve at P, the value for n will increase.
53
5.4.2.2 Other models
5.4.2.2.1
The model of Brooks and Corey
The model of Brooks and Corey (1964) can be expressed as
*;
(19)
where e = porosity (m3-m"3),
hb = bubbling pressure or air entry value (cm HbO),
X = pore size index (-).
The model of Brooks and Corey, and the model of Campbell (see 5.4.2.2.2.) do not permit a
representation of the total soil moisture retention curve - they only represent that part of the
curve for pressure heads lower than the air entry value -, whereas the model of van
Genuchten does so. The model of van Genuchten does not account for the air entry value but
does have an inflection point, allowing the van Genuchten model to perform better than the
Brooks and Corey, and the Campbell model for many soils, particularly for data near
saturation.
5.4.2.2.2
The model of Campbell
The model of Campbell (1974) can be represented by :
h
(20)
where H^ - bubbling pressure or air entry value (cm H2O),
b = constant.
5.4.3
Methods to obtain the pF curve without laboratory determinations
5.4.3.1 Soil texture reference curves
The simplest method for estimating the h{6) relationship is to use soil texture reference
curves. These curves give the soil water characteristic curve for several soil textural classes
(see figure 27).
54
~1
-10
-100
-1O0O
-16.000 -100,000
-100
-*000
-10,000
-100,000
uunc tx*m& h*** fern*
-too
-1000 -*too& -100.600
Marie po<em^ N a d fcm|
Figure 27 Soil water retention curves for USDA soil textures (Rawls et al., 1993)
5.4.3.2 Estimating soil water retention curves from soil physical characteristics
5.4.3.2.1
Introduction
The amount of water retained at high soil water potentials (> -10 kPa) depends primarily on
the pore size distribution, and is thus strongly affected by soil structure, bulk density, and
porosity. As the soil dries out (< -150 kPa), water adsorption becomes critical, and those soil
properties that affect specific surface become important such as texture, organic matter
content, and clay mineralogy. Most of the models that are developed to estimate the soil
moisture characteristic curve from soil physical characteristics, the so called pedotransfer
functions, make use of one or more of those characteristics.
Two approaches can be distinguished for estimating soil water retention characteristics from
soil properties. The first approach estimates soil water retention values from soil physical
properties using regression analysis. The second approach estimates parameters for water
retention models (like van Genuchten, Brooks and Corey, and Campbell) from soil physical
properties using regression analysis. An overview of both approaches is given in Rawls et al.
(1991) and van Genuchten et al. (1992). More recently Kern (1995) evaluated six of the most
cited models that use soil physical characteristics. These include the models of Gupta and
Larson (1979), Rawls et al. (1982, 1983), De Jong et al. (1983), Cosby et al. (1984), Saxton
et al. (1986) and Vereecken et al. (1989).
55
5.4.3.2.2
Estimation of specific points on the soil water retention curve
The most cited models to relate soil properties to the soil moisture held at specific matric
potentials are those of Gupta and Larson (1979) and Rawls et al. (1982, 1983). To increase
the accuracy of the regression equations using physical soil properties, Rawls et al.
introduced the moisture content held at -1500 kPa (pF 4.2) or at both -33 kPa (pF 2.54) and
-1500 kPa. Adding these variables, which require more costly and/or time-consuming
laboratory procedures, increase the explained variation from 76% to 95%. The model of
Rawls et al. can thus be written as :
6X = a + b - sand + c • silt + d - clay + e • O.M. + / • pb + g • 0O|33bar
+
h ' #15bar
(21)
where 0X
= predicted soil moisture content for a given suction x (cm"3-cm"3),
sand = percentage sand (%),
silt
= percentage silt (%),
clay = percentage clay (%),
O.M. = percentage organic matter (%),
pb
= bulk density (g-cm"3),
0o,33 bar = moisture content at a matric potential of - 33 kPa (cm"3-cm~3),
Ois bar ~ moisture content at a matric potential of - 1500 kPa (cm~3-cm~3),
aXoXh - regression coefficients from tables.
Thus, in its simplest form the model of Rawls et al. exists of the first 6 terms of equation (21).
The values for regression coefficients differ depending on the chosen simplicity of the model
(6, 7, or 8 terms) and the given matric potential. The tables of Rawls et al. include 12
different matric potentials, and hence a set of 12 discrete points can be determined. A best
fitting curve can then be drawn through this set of points by modelling (e.g. van Genuchten).
The model of Gupta and Larson is similar with this difference that the intercept a is equal to
zero (and hence the regression coefficients are different).
5.4.3.2.3
Estimation of soil water retention model parameters
With the increased interest on modelling, the need for a continuous function describing the
soil water retention curve has become very important. Frequently water retention data is fitted
to a water retention model (see 5.4.2) and the model parameters subsequently related to
physical soil properties using regression analysis.
56
One approach to relate the model parameters to soil properties is to develop average
parameter values as a function of soil texture classes. Tables are available with values for the
parameters of the different models. Clapp en Hornberger (1978), De Jong (1982) and Raws et
al. (1982) calculated, for the different USDA textural classes, values for the models of
Brooks and Corey, and Campbell. Vereecken et al. (1989) calculated the values for the
parameters of the van Genuchten models for the seven textural classes of the Belgian textural
triangle (see table 2).
The second approach uses linear regression to relate the model parameters to soil physical
properties. Examples of this approach are the pedotransfer functions of Bloemen (1977),
Cosby et al. (1984) and Rawls and Brakensiek (1985) for the Brooks and Corey model. Using
the correspondence between the model parameters given in 2, i.e. Hb = h^ b = 1//1, a = hb~l
and n = X +1, the same pedotransfer functions can be used with the Campbell and the van
Genuchten model. Bloemen developed empirical expression relating the model parameters to
a particle size distribution index and the median of particle size, whereas in Cosby et al. the
parameters are function of percentage clay and sand. Rawls and Brakensiek included also
porosity. Saxton et al. (1986) used the percentage clay and sand to calculate the parameters of
a model that was derived from the model of Campbell. Widely used pedotransfer functions
for the model of van Genuchten are those of Vereecken et al. (1989) which were developed
based on the physical characteristics of 182 horizons of 40 different Belgian soils :
6S = 0.81 - 0.283 • pb + 0.001 • clay
6r = 0.015 + 0.005 • clay + 0.014 • C
ln(a) = - 2.486 + 0.025 • sand - 0.351 • C - 2.617 • pb - 0.023 • clay
ln(7i) = 0.053 - 0.009'sand - 0.013-clay + 0.00015-sand2
where
C = carbon content (%).
De Jong et al. (1983) followed a similar approach but calculated the parameters from
percentage organic matter, silt and clay. These parameters are then substituted in a model that
expresses gravimetric moisture content as a function of matric potential by linear regression.
A third more recent approach introduced by Tyler and Wheatcraft (1990) uses fractal
mathematics and scaled similarities to show that the empirical constant in the Arya and Paris
(1981) model is equivalent to the fractal dimension of the tortuous fractal pore. This
information can be used to predict soil water retention from measured particle size
distributions.
57
5.4.4
Determination of the unsaturated hydraulic conductivity K(6) by modelling
5.4.4.1 Introduction
To determine the unsaturated hydraulic conductivity K(0), several approaches can be
distinguished. The most common technique is to determine K{9) from the pore-size
distribution and the saturated hydraulic conductivity Ks. The pore-size distribution can be
derived from the soil moisture characteristic curve. The Ks has to be determined in the
laboratory or in the field, or has to be calculated based on soil physical properties.
The K(0) relationship can also be determined from readily available soil physical
characteristics.
5.4.4.2 Estimating the unsaturated hydraulic conductivity K(6) from water retention data
5.4.4.2.1
Introduction
A popular alternative to direct measurement of K(0) is the use of theoretical methods which
predict K{9) from more easily determined field or laboratory water retention data. Theoretical
methods are usually based on statistical pore-size distribution models which assume water
flow through cylindrical pores, and incorporate the equations of Darcy and Poiseuille. Three
broad groups of models predicting the unsaturated hydraulic conductivity from measured
water retention data can be identified, i.e., the models proposed by Childs and Collis-George
(1950), Burdine (1953) and Mualem (1976).
58
Table 2
Mean, minimum and maximum value for the parameters of the van Genuchten
model for the seven textural classes according to the Belgian classification (N is the number of
observations) with m = 1
Textural class
Heavy clay (U)
N
ft
ft
a-10-3
n
Minimum
Maximum
0.55
0.27
1.60
0.66
0.52
0.25
0.60
0.57
0.57
0.29
2.40
0.71
0.44
0.16
2.00
0.63
0.39
0.00
0.49
0.30
7.77
0.96
0.42
0.11
0.46
0.14
1.56
0.80
0.38
0.04
0.40
0.42
0.41
0.09
2.29
0.86
0.36
0.00
0.20
0.32
0.48
0.22
5.70
1.23
0.37
0.09
2.50
0.92
0.35
0.07
0.61
0.58
0.42
0.11
5.60
1.17
0.41
0.09
7.42
1.21
0.32
0.07
3.41
0.76
0.53
0.14
20.61
1.27
0.39
0.04
12.30
1.68
0.30
0.00
1.60
0.69
0.54
0.10
37.10
2.62
10
Clay (E)
ft
a
a-10-3
n
2.00
0.30
33
Loam (A)
ft
ft
Sandy loam (L)
aa- 1O-3
n
Light sandy loam (P)
Loamy sand (S)
ft
10
10
a
3
tf-1071
Sand (Z)
9S
ft
a-10-3
4.20
1.44
55
A
71
Mean
62
59
5.4.4.2.2
Estimating K(0) based on the model of Childs and Collis-George
The original equation by Childs and Collis-George (1950) has been later modified by
Marshall (1958), Millington en Quirk (1959) and Jackson (1972). Jackson formulated the
unsaturated hydraulic conductivity at moisture content 9t ,K(9j), as :
m
P
f
"
(22)
•'-
saturated hydraulic conductivity (m-s"1),
amount of pores filled with water = (9/ (cm3-cm"3),
pressure head at midpoint of each 0 increment./ = / (m),
number of moisture content class where i =1,2, ... m, and / = 1 at saturation and
m = the total number of increments,
p = arbitrary constant assigned values of 0 to 4/3 by various workers (a value of unity
was found to be satisfactory by Jackson) (-).
where Ks
sj
hj
i
=
=
=
=
The saturated hydraulic conductivity Ks can be derived from field or laboratory measurements,
or can be calculated based on statistical pore-size distribution models. Childs and
Collis-George (1950) defined Ks as :
r)
Ks =
A0 • 5 > f 2
1
21PgT
(23)
surface tension of water (N-m"1),
viscosity of water (kg-m'^s"1),
density of water (kg-m"3),
gravitational acceleration (m-s"2),
tortuosity (m-m"1),
0 increment (m3-m"3),
number of different capillary size classes in the bundle of tubes making up the soil
column,
hi = pressure head at midpoint of each 9 increment / (m).
where y =
r/ =
pw =
g =
T =
A0=
n =
60
Marshall (1958) and Millington and Quirk (1959) defined Ks as
K. =
2-tj-pw-g
r,P
III
n<
f.
(24)
5.4.4.2.3 Estimating K(9) based on the model of Burdine
Burdine's (1953) model was later applied by Brooks and Corey (1964) to derive their
classical function for the unsaturated hydraulic conductivity K(9):
(25)
K(ff) =
£ - Or
where n — 3 + 21X.
5AA.2A
Estimating K(9) based on the model of Mualem
Van Genuchten (1980) used the model of Mualem (1976) to derive his well-known equation
for unsaturated hydraulic conductivity K(6) or K{h):
\ /n
K(6) = Ks
e - er 12
os - or
1 -
0 - 6r \m
ds - 9r
-m
i-l
K(h) = Ks
n
l+(a.\h\)
where a, n and m
m
\2
= parameters from the van Genuchten model.
61
(26)
(27)
5.4.4.3 Estimating the unsaturated hydraulic conductivity K(0) from soil physical
characteristics
5.4.4.3.1
Soil texture reference curves
As for the soil moisture retention curve, the X-<9 relationship can be derived from soil texture
reference curves (see figure 28) The saturated hydraulic conductivity Ks can also be derived
from soil texture reference curves (see figure 29).
Sami
to1
Loamy sand
Sandy loam
loam
Silt loam
Sandy day loam
Clay loam
Silly day bam
Sandy day
Stfty day
Clay
10°
tcr
1
I to-3 .
ttr
t<r5
0
10
20
30
40
50
60
70
80
90 100
% Saturation
Figure 28
Hydraulic conductivity sorted by soil texture (Rawls et aL, 1982)
Figure 29
Saturated hydraulic conductivity Ks for USDA soil texture triangle
(Rawls and Brakensiek, 1985)
62
5.4.4.3.2
Pedotransfer functions to estimate K(6)
Very popular pedotransfer functions to determine Ks and K{6) or K(y/) are those proposed by
Campbell (1985):
1.3-5
(-6'9'mc-3-7'ms)
'6
OKO IA-3
•O.OO-IU
K(6) = KSA^\
K{¥) = Ks-\^\
(29)
"
where A^5 = saturated hydraulic conductivity (cm-h"1),
mc = clay mass fraction,
ms = silt mass fraction,
y/es = air entry potential (J-kg"1) for a standard bulk density of 1,300 kg-m"3,
—
-
f\ K ^7-05
-Q.5-ag
,
t/g = geometric mean particle diameter,
]T mi • In d(
<yg = geometric standard deviation,
==
^v
mt n l
"h
?
m,- = mass fraction of textural class /,
J/ = arithmetic mean diameter of class /,
m
= 2-5 + 3,
y/e = air entry potential (J-kg"1),
0.67-5
63
(0Q\
\A*)
(30)
5.4.5
References
Arya, L.M. and J.F. Paris (1981). A physicoemperical model to predict the soil moisture
characteristic from particle-size distribution and bulk density data. Soil. Sci. Soc. Am. J., 45,
1023-1030.
Bloemen, G.W. (1980). Calculation of capillary conductivity and capillary rise from grain
size distribution. I. Real and theoretical values of the exponent in a formula of Brooks and
Corey for the calculation of hydraulic conductivities. ICW Wageningen nota no. 952. II.
Assessment of the values of the exponent in a formula of Brooks and Corey for the
calculation of hydraulic conductivity from grain size distribution. ICW Wageningen nota no.
962. III. Air entry pressure and saturated conductivity calculated from grain size distribution
and median grain size. ICW Wageningen nota no. 990. IV. Capillary rise in soil types and
soil profiles. ICW Wageningen nota no. 1013.
Brooks, R.H. and A.T. Corey (1964). Hydraulic properties of porous media. Hydrology
Paper 3, Colorado State University, Fort Collins, Colorado, USA.
Burdine, N.T. (1952). Relative permeability calculations for pore-size distribution data.
Trans. AIME, 198, 34-42.
Campbell, G.S. (1974). A simple method for determining unsaturated conductivity from
moisture retention data. Soil Sci., 117, 311-314.
Campbell, G.S. (1985). Soil physics with BASIC. Transport models for soil-plant systems.
Developments in Soil Science 14. Dept. of Agronomy and Soils, Washington State
University, Pullman, USA.
Childs, E.C. and N. Collis-George (1950). The permeability of porous materials. Proc. R.
Soc. London Ser. A, 201, 392-405.
Clapp, R.B. and G.M. Hornberger (1978). Empirical equations for some hydraulic
properties. Water Resour. Res., 15, 601-604.
Cosby, B.J., G.M. Hornberger, R.B. Clapp and T.R. Ginn (1984). A statistical exploration
of soil moisture characteristics to the physical properties of soils. Water Resour. Res., 20,
682-690.
De Jong, R. (1982). Assessment of empirical parameters that describe soil water
characteristics. Can. Agric. Engin., 24, 65-70.
64
De Jong, R., C.A. Campbell and W. Nicholaichuk (1983). Water retention equations and
their relationship to soil organic matter and particle size distributions for disturbed samples.
Can. J. Soil Sci., 63, 291-302.
Gupta, S.C. and W.E. Larson (1979). Estimating soil water retention characteristics from
particle size distribution, organic matter percent and bulk density. Water Resour. Res., 15,
1633-1635.
Jackson, R.A. (1972). On the calculation of hydraulic conductivity. Soil Sci. Soc. Am. Proc,
36, 380-383.
Kern, J.S. (1995). Evaluation of soil water retention models based on soil physical
properties. Soil Sci. Soc. Am. J., 59, 1134-1141.
Marquardt, D.W. (1963). An algorithm for least-squares estimation of non-linear
parameters. J. Soc. Ind. Appl. Math., 11, 431-441.
Marshall, T.J. (1958). A Relation between permeability and size distribution of pores. J. Soil
Sci., 9, 1-8.
Millington, R.J. and J.P. Quirk (1959). Permeability of porous media. Nature, 183,
387-388.
Mualem, Y. (1976). A new model for predicting the hydraulic conductivity of unsaturated
porous media. Water Resour. Res., 12, 513-522.
Rawls, W.J., D.L. Brakensiek and K.E. Saxton (1982). Estimation of soil water properties.
Trans. ASAE, 25, 1316-1328.
Rawls, W.J., D.L. Brakensiek and B.Soni (1983). Agricultural management effects on soil
water processes, Part I. Soil water retention and Green and Ampt infiltration parameters.
Trans. ASAE, 26, 1747-1752.
Rawls, W.J. and D.L. Brakensiek (1985). Prediction of soil water properties for
hydrological modelling. In : E.B. Jones en T.J. Wards (eds.). Watershed management in the
eighties. Proc. of Symp. Sponsored by Comm. on Watershed Management, I & D Division,
ASCE. ASCE Convention, Denver, Colorado, USA, April 30-May 1, 293-299.
Rawls, W.J., T.J. Gish and D.L. Brakensiek (1991). Estimating soil water retention from
soil physical properties and characteristics. Adv. Soil Sci., 16, 213-234.
65
Rawls, W.J., L.R. Ahuja, D.L. Brakensiek and A. Shirmohammadi (1993). Infiltration
and soil water movement. In : D.R. Maidment (ed.). Handbook of hydrology. McGraw-Hill,
Inc., USA.
Saxton, K.E., W.J. Rawls, J.S. Romberger and R.L Papendick (1986). Estimating
generalised soil-water characteristics from texture. Soil Sci. Soc. Am. J., 50, 1031-1036.
Tyler, S.W. and S.W. Wheatcraft (1990). Fractal processes in soil water retention. Water
Resour. Res., 26, 1047-1054.
van Genuchten, M.Th. (1978). Calculating the unsaturated hydraulic conductivity with a
new closed-form analytical model. Research Report 78-WR-08, Water Research Program,
Dept. of Civil Engineering, Princeton University, Princeton, New Jersey, USA.
van Genuchten, M. Th. (1980). A closed-form equation for predicting the hydraulic
conductivity of unsaturated soils. Soil Sci. Soc. Am. J., 44, 892-898.
van Genuchten, M.Th., F.J. Leij and S.R. Yates (1991). The RETC code for quantifying
the hydraulic functions of unsaturated soils. EPA/600/2-91/065, R.S. Kerr Environmental
Research Laboratory, Office of Research and Development, U.S. Environmental Protection
Agency, ADA, Oklahoma, USA.
van Genuchten, M.Th., F.J. Leij and L.J. Lund (1992). Indirect methods for estimating the
hydraulic properties of unsaturated soils. Univ. of California, Riverside, USA.
Vereecken, H., J. Maes, J. Feyen and P. Darius (1989). Estimating the soil moisture
retention characteristic from texture, bulk density and carbon content. Soil Sci., 148, 389-403.
Yates, S.R., M.Th. van Genuchten, A.W. Warrick and F.J. Leij (1992). Analysis of
measured, predicted, and estimated hydraulic conductivity using the RETC computer
program. Soil Sci. Soc. Am. J., 56, 347-354.
66